Showing 1–46 of 46 results for author: Patanè, A

Uncertainty Propagation in Stochastic Systems via Mixture Models with Error Quantification

Authors: Eduardo Figueiredo, Andrea Patane, Morteza Lahijanian, Luca Laurenti

Abstract: Uncertainty propagation in non-linear dynamical systems has become a key problem in various fields including control theory and machine learning. In this work we focus on discrete-time non-linear stochastic dynamical systems. We present a novel approach to approximate the distribution of the system over a given finite time horizon with a mixture of distributions. The key novelty of our approach is… ▽ More Uncertainty propagation in non-linear dynamical systems has become a key problem in various fields including control theory and machine learning. In this work we focus on discrete-time non-linear stochastic dynamical systems. We present a novel approach to approximate the distribution of the system over a given finite time horizon with a mixture of distributions. The key novelty of our approach is that it not only provides tractable approximations for the distribution of a non-linear stochastic system, but also comes with formal guarantees of correctness. In particular, we consider the total variation (TV) distance to quantify the distance between two distributions and derive an upper bound on the TV between the distribution of the original system and the approximating mixture distribution derived with our framework. We show that in various cases of interest, including in the case of Gaussian noise, the resulting bound can be efficiently computed in closed form. This allows us to quantify the correctness of the approximation and to optimize the parameters of the resulting mixture distribution to minimize such distance. The effectiveness of our approach is illustrated on several benchmarks from the control community. △ Less

Submitted 22 March, 2024; originally announced March 2024.

Probabilistic Reach-Avoid for Bayesian Neural Networks

Authors: Matthew Wicker, Luca Laurenti, Andrea Patane, Nicola Paoletti, Alessandro Abate, Marta Kwiatkowska

Abstract: Model-based reinforcement learning seeks to simultaneously learn the dynamics of an unknown stochastic environment and synthesise an optimal policy for acting in it. Ensuring the safety and robustness of sequential decisions made through a policy in such an environment is a key challenge for policies intended for safety-critical scenarios. In this work, we investigate two complementary problems: f… ▽ More Model-based reinforcement learning seeks to simultaneously learn the dynamics of an unknown stochastic environment and synthesise an optimal policy for acting in it. Ensuring the safety and robustness of sequential decisions made through a policy in such an environment is a key challenge for policies intended for safety-critical scenarios. In this work, we investigate two complementary problems: first, computing reach-avoid probabilities for iterative predictions made with dynamical models, with dynamics described by Bayesian neural network (BNN); second, synthesising control policies that are optimal with respect to a given reach-avoid specification (reaching a "target" state, while avoiding a set of "unsafe" states) and a learned BNN model. Our solution leverages interval propagation and backward recursion techniques to compute lower bounds for the probability that a policy's sequence of actions leads to satisfying the reach-avoid specification. Such computed lower bounds provide safety certification for the given policy and BNN model. We then introduce control synthesis algorithms to derive policies maximizing said lower bounds on the safety probability. We demonstrate the effectiveness of our method on a series of control benchmarks characterized by learned BNN dynamics models. On our most challenging benchmark, compared to purely data-driven policies the optimal synthesis algorithm is able to provide more than a four-fold increase in the number of certifiable states and more than a three-fold increase in the average guaranteed reach-avoid probability. △ Less

Submitted 3 October, 2023; originally announced October 2023.

Comments: 47 pages, 10 figures. arXiv admin note: text overlap with arXiv:2105.10134

A magnetically-induced Coulomb gap in graphene due to electron-electron interactions

Authors: E. E. Vdovin, M. T. Greenaway, Yu. N. Khanin, S. V. Morozov, O. Makarovsky, A. Patanè, A. Mishchenko, S. Slizovskiy, V. I. Fal'ko, A. K. Geim, K. S. Novoselov, L. Eaves

Abstract: Insights into the fundamental properties of graphene's Dirac-Weyl fermions have emerged from studies of electron tunnelling transistors in which an atomically thin layer of hexagonal boron nitride (hBN) is sandwiched between two layers of high purity graphene. Here, we show that when a single defect is present within the hBN tunnel barrier, it can inject electrons into the graphene layers and its… ▽ More Insights into the fundamental properties of graphene's Dirac-Weyl fermions have emerged from studies of electron tunnelling transistors in which an atomically thin layer of hexagonal boron nitride (hBN) is sandwiched between two layers of high purity graphene. Here, we show that when a single defect is present within the hBN tunnel barrier, it can inject electrons into the graphene layers and its sharply defined energy level acts as a high resolution spectroscopic probe of electron-electron interactions in graphene. We report a magnetic field dependent suppression of the tunnel current flowing through a single defect below temperatures of $\sim$ 2 K. This is attributed to the formation of a magnetically-induced Coulomb gap in the spectral density of electrons tunnelling into graphene due to electron-electron interactions. △ Less

Submitted 4 July, 2023; originally announced July 2023.

Journal ref: Communications Physics volume 6, Article number: 159 (2023)

Adversarial Robustness Certification for Bayesian Neural Networks

Authors: Matthew Wicker, Andrea Patane, Luca Laurenti, Marta Kwiatkowska

Abstract: We study the problem of certifying the robustness of Bayesian neural networks (BNNs) to adversarial input perturbations. Given a compact set of input points $T \subseteq \mathbb{R}^m$ and a set of output points $S \subseteq \mathbb{R}^n$, we define two notions of robustness for BNNs in an adversarial setting: probabilistic robustness and decision robustness. Probabilistic robustness is the probabi… ▽ More We study the problem of certifying the robustness of Bayesian neural networks (BNNs) to adversarial input perturbations. Given a compact set of input points $T \subseteq \mathbb{R}^m$ and a set of output points $S \subseteq \mathbb{R}^n$, we define two notions of robustness for BNNs in an adversarial setting: probabilistic robustness and decision robustness. Probabilistic robustness is the probability that for all points in $T$ the output of a BNN sampled from the posterior is in $S$. On the other hand, decision robustness considers the optimal decision of a BNN and checks if for all points in $T$ the optimal decision of the BNN for a given loss function lies within the output set $S$. Although exact computation of these robustness properties is challenging due to the probabilistic and non-convex nature of BNNs, we present a unified computational framework for efficiently and formally bounding them. Our approach is based on weight interval sampling, integration, and bound propagation techniques, and can be applied to BNNs with a large number of parameters, and independently of the (approximate) inference method employed to train the BNN. We evaluate the effectiveness of our methods on various regression and classification tasks, including an industrial regression benchmark, MNIST, traffic sign recognition, and airborne collision avoidance, and demonstrate that our approach enables certification of robustness and uncertainty of BNN predictions. △ Less

Submitted 23 June, 2023; originally announced June 2023.

BNN-DP: Robustness Certification of Bayesian Neural Networks via Dynamic Programming

Authors: Steven Adams, Andrea Patane, Morteza Lahijanian, Luca Laurenti

Abstract: In this paper, we introduce BNN-DP, an efficient algorithmic framework for analysis of adversarial robustness of Bayesian Neural Networks (BNNs). Given a compact set of input points $T\subset \mathbb{R}^n$, BNN-DP computes lower and upper bounds on the BNN's predictions for all the points in $T$. The framework is based on an interpretation of BNNs as stochastic dynamical systems, which enables the… ▽ More In this paper, we introduce BNN-DP, an efficient algorithmic framework for analysis of adversarial robustness of Bayesian Neural Networks (BNNs). Given a compact set of input points $T\subset \mathbb{R}^n$, BNN-DP computes lower and upper bounds on the BNN's predictions for all the points in $T$. The framework is based on an interpretation of BNNs as stochastic dynamical systems, which enables the use of Dynamic Programming (DP) algorithms to bound the prediction range along the layers of the network. Specifically, the method uses bound propagation techniques and convex relaxations to derive a backward recursion procedure to over-approximate the prediction range of the BNN with piecewise affine functions. The algorithm is general and can handle both regression and classification tasks. On a set of experiments on various regression and classification tasks and BNN architectures, we show that BNN-DP outperforms state-of-the-art methods by up to four orders of magnitude in both tightness of the bounds and computational efficiency. △ Less

Submitted 19 June, 2023; originally announced June 2023.

Comments: To appear at ICML 2023

Room temperature quantum Hall effect in a gated ferroelectric-graphene heterostructure

Authors: Anubhab Dey, Nathan Cottam, Oleg Makarovskiy, Wen**g Yan, Vaidotas Mišeikis, Camilla Coletti, James Kerfoot, Vladimir Korolkov, Laurence Eaves, Jasper F. Linnartz, Arwin Kool, Steffen Wiedmann, Amalia Patanè

Abstract: The quantum Hall effect is widely used for the investigation of fundamental phenomena, ranging from topological phases to composite fermions. In particular, the discovery of a room temperature resistance quantum in graphene is significant for compact resistance standards that can operate above cryogenic temperatures. However, this requires large magnetic fields that are accessible only in a few hi… ▽ More The quantum Hall effect is widely used for the investigation of fundamental phenomena, ranging from topological phases to composite fermions. In particular, the discovery of a room temperature resistance quantum in graphene is significant for compact resistance standards that can operate above cryogenic temperatures. However, this requires large magnetic fields that are accessible only in a few high magnetic field facilities. Here, we report on the quantum Hall effect in graphene encapsulated by the ferroelectric insulator CuInP2S6. Electrostatic gating of the graphene channel enables the Fermi energy to be tuned so that electrons in the localized states of the insulator are in equilibrium with the current-carrying, delocalized states of graphene. Due to the presence of strongly bound states in this hybrid system, a quantum Hall plateau can be achieved at room temperature in relatively modest magnetic fields. This phenomenon offers the prospect for the controlled manipulation of the quantum Hall effect at room temperature. △ Less

Submitted 26 May, 2023; originally announced May 2023.

Comments: 24 pages, 6 figures

Individual Fairness in Bayesian Neural Networks

Authors: Alice Doherty, Matthew Wicker, Luca Laurenti, Andrea Patane

Abstract: We study Individual Fairness (IF) for Bayesian neural networks (BNNs). Specifically, we consider the $ε$-$δ$-individual fairness notion, which requires that, for any pair of input points that are $ε$-similar according to a given similarity metrics, the output of the BNN is within a given tolerance $δ>0.$ We leverage bounds on statistical sampling over the input space and the relationship between a… ▽ More We study Individual Fairness (IF) for Bayesian neural networks (BNNs). Specifically, we consider the $ε$-$δ$-individual fairness notion, which requires that, for any pair of input points that are $ε$-similar according to a given similarity metrics, the output of the BNN is within a given tolerance $δ>0.$ We leverage bounds on statistical sampling over the input space and the relationship between adversarial robustness and individual fairness to derive a framework for the systematic estimation of $ε$-$δ$-IF, designing Fair-FGSM and Fair-PGD as global,fairness-aware extensions to gradient-based attacks for BNNs. We empirically study IF of a variety of approximately inferred BNNs with different architectures on fairness benchmarks, and compare against deterministic models learnt using frequentist techniques. Interestingly, we find that BNNs trained by means of approximate Bayesian inference consistently tend to be markedly more individually fair than their deterministic counterparts. △ Less

Submitted 21 April, 2023; originally announced April 2023.

On the Robustness of Bayesian Neural Networks to Adversarial Attacks

Authors: Luca Bortolussi, Ginevra Carbone, Luca Laurenti, Andrea Patane, Guido Sanguinetti, Matthew Wicker

Abstract: Vulnerability to adversarial attacks is one of the principal hurdles to the adoption of deep learning in safety-critical applications. Despite significant efforts, both practical and theoretical, training deep learning models robust to adversarial attacks is still an open problem. In this paper, we analyse the geometry of adversarial attacks in the large-data, overparameterized limit for Bayesian… ▽ More Vulnerability to adversarial attacks is one of the principal hurdles to the adoption of deep learning in safety-critical applications. Despite significant efforts, both practical and theoretical, training deep learning models robust to adversarial attacks is still an open problem. In this paper, we analyse the geometry of adversarial attacks in the large-data, overparameterized limit for Bayesian Neural Networks (BNNs). We show that, in the limit, vulnerability to gradient-based attacks arises as a result of degeneracy in the data distribution, i.e., when the data lies on a lower-dimensional submanifold of the ambient space. As a direct consequence, we demonstrate that in this limit BNN posteriors are robust to gradient-based adversarial attacks. Crucially, we prove that the expected gradient of the loss with respect to the BNN posterior distribution is vanishing, even when each neural network sampled from the posterior is vulnerable to gradient-based attacks. Experimental results on the MNIST, Fashion MNIST, and half moons datasets, representing the finite data regime, with BNNs trained with Hamiltonian Monte Carlo and Variational Inference, support this line of arguments, showing that BNNs can display both high accuracy on clean data and robustness to both gradient-based and gradient-free based adversarial attacks. △ Less

Submitted 28 February, 2024; v1 submitted 13 July, 2022; originally announced July 2022.

Comments: arXiv admin note: text overlap with arXiv:2002.04359

Individual Fairness Guarantees for Neural Networks

Authors: Elias Benussi, Andrea Patane, Matthew Wicker, Luca Laurenti, Marta Kwiatkowska

Abstract: We consider the problem of certifying the individual fairness (IF) of feed-forward neural networks (NNs). In particular, we work with the $ε$-$δ$-IF formulation, which, given a NN and a similarity metric learnt from data, requires that the output difference between any pair of $ε$-similar individuals is bounded by a maximum decision tolerance $δ\geq 0$. Working with a range of metrics, including t… ▽ More We consider the problem of certifying the individual fairness (IF) of feed-forward neural networks (NNs). In particular, we work with the $ε$-$δ$-IF formulation, which, given a NN and a similarity metric learnt from data, requires that the output difference between any pair of $ε$-similar individuals is bounded by a maximum decision tolerance $δ\geq 0$. Working with a range of metrics, including the Mahalanobis distance, we propose a method to overapproximate the resulting optimisation problem using piecewise-linear functions to lower and upper bound the NN's non-linearities globally over the input space. We encode this computation as the solution of a Mixed-Integer Linear Programming problem and demonstrate that it can be used to compute IF guarantees on four datasets widely used for fairness benchmarking. We show how this formulation can be used to encourage models' fairness at training time by modifying the NN loss, and empirically confirm our approach yields NNs that are orders of magnitude fairer than state-of-the-art methods. △ Less

Submitted 11 May, 2022; originally announced May 2022.

Terahertz control of photoluminescence emission in few-layer InSe

Authors: Tommaso Venanzi, Malte Selig, Alexej Pashkin, Stephan Winnerl, Manuel Katzer, Himani Arora, Artur Erbe, Amalia Patanè, Zakhar R. Kudrynskyi, Zakhar D. Kovalyuk, Leonetta Baldassarre, Andreas Knorr, Manfred Helm, Harald Schneider

Abstract: A promising route for the development of opto-elelctronic technology is to use terahertz radiation to modulate the optical properties of semiconductors. Here we demonstrate the dynamical control of photoluminescence (PL) emission in few-layer InSe using picosecond terahertz pulses. We observe a strong PL quenching (up to 50%) after the arrival of the terahertz pulse followed by a reversible recove… ▽ More A promising route for the development of opto-elelctronic technology is to use terahertz radiation to modulate the optical properties of semiconductors. Here we demonstrate the dynamical control of photoluminescence (PL) emission in few-layer InSe using picosecond terahertz pulses. We observe a strong PL quenching (up to 50%) after the arrival of the terahertz pulse followed by a reversible recovery of the emission on the time scale of 50ps at T =10K. Microscopic calculations reveal that the origin of the photoluminescence quenching is the terahertz absorption by photo-excited carriers: this leads to a heating of the carriers and a broadening of their distribution, which reduces the probability of bimolecular electron-hole recombination and, therefore, the luminescence. By numerically evaluating the Boltzmann equation, we are able to clarify the individual roles of optical and acoustic phonons in the subsequent cooling process. The same PL quenchingmechanismis expected in other van derWaals semiconductors and the effectwill be particularly strong for materials with low carrier masses and long carrier relaxation time, which is the case for InSe. This work gives a solid background for the development of opto-electronic applications based on InSe, such as THz detectors and optical modulators. △ Less

Submitted 18 February, 2022; originally announced February 2022.

Comments: The following article has been accepted by Applied Physics Letters. After it is published, it will be found at https://publishing.aip.org/resources/librarians/products/journals/

Design and characterization of effective solar cells

Authors: Varun Ojha, Giorgio Jansen, Andrea Patane, Antonino La Magna, Vittorio Romano, Giuseppe Nicosia

Abstract: We propose a two-stage multi-objective optimization framework for full scheme solar cell structure design and characterization, cost minimization and quantum efficiency maximization. We evaluated structures of 15 different cell designs simulated by varying material types and photodiode do** strategies. At first, non-dominated sorting genetic algorithm~II (NSGA-II) produced Pareto-optimal-solutio… ▽ More We propose a two-stage multi-objective optimization framework for full scheme solar cell structure design and characterization, cost minimization and quantum efficiency maximization. We evaluated structures of 15 different cell designs simulated by varying material types and photodiode do** strategies. At first, non-dominated sorting genetic algorithm~II (NSGA-II) produced Pareto-optimal-solutions sets for respective cell designs. Then, on investigating quantum efficiencies of all cell designs produced by NSGA-II, we applied a new multi-objective optimization algorithm~II (OptIA-II) to discover the Pareto fronts of select (three) best cell designs. Our designed OptIA-II algorithm improved the quantum efficiencies of all select cell designs and reduced their fabrication costs. We observed that the cell design comprising an optimally doped zinc-oxide-based transparent conductive oxide (TCO) layer and rough silver back reflector (BR) offered a quantum efficiency ($Q_e$) of $0.6031.$ Overall, this paper provides a full characterization of cell structure designs. It derives a relationship between quantum efficiency, $Q_e$ of a cell with its TCO layer's do** methods and TCO and BR layer's material types. Our solar cells design characterization enables us to perform a cost-benefit analysis of solar cells usage in real-world applications △ Less

Submitted 11 September, 2021; originally announced September 2021.

Comments: published in 2021

Journal ref: Energy Systems, 2021

Tunable spin-orbit coupling in two-dimensional InSe

Authors: A. Ceferino, S. J. Magorrian, V. Zólyomi, D. A. Bandurin, A. K. Geim, A. Patanè, Z. D. Kovalyuk, Z. R. Kudrynskyi, I. V. Grigorieva, V. I. Fal'ko

Abstract: We demonstrate that spin-orbit coupling (SOC) strength for electrons near the conduction band edge in few-layer $γ$-InSe films can be tuned over a wide range. This tunability is the result of a competition between film-thickness-dependent intrinsic and electric-field-induced SOC, potentially, allowing for electrically switchable spintronic devices. Using a hybrid $\mathbf{k\cdot p}$ tight-binding… ▽ More We demonstrate that spin-orbit coupling (SOC) strength for electrons near the conduction band edge in few-layer $γ$-InSe films can be tuned over a wide range. This tunability is the result of a competition between film-thickness-dependent intrinsic and electric-field-induced SOC, potentially, allowing for electrically switchable spintronic devices. Using a hybrid $\mathbf{k\cdot p}$ tight-binding model, fully parameterized with the help of density functional theory computations, we quantify SOC strength for various geometries of InSe-based field-effect transistors. The theoretically computed SOC strengths are compared with the results of weak antilocalization measurements on dual-gated multilayer InSe films, interpreted in terms of Dyakonov-Perel spin relaxation due to SOC, showing a good agreement between theory and experiment. △ Less

Submitted 8 June, 2021; originally announced June 2021.

Journal ref: Phys. Rev. B 104, 125432 (2021)

Certification of Iterative Predictions in Bayesian Neural Networks

Authors: Matthew Wicker, Luca Laurenti, Andrea Patane, Nicola Paoletti, Alessandro Abate, Marta Kwiatkowska

Abstract: We consider the problem of computing reach-avoid probabilities for iterative predictions made with Bayesian neural network (BNN) models. Specifically, we leverage bound propagation techniques and backward recursion to compute lower bounds for the probability that trajectories of the BNN model reach a given set of states while avoiding a set of unsafe states. We use the lower bounds in the context… ▽ More We consider the problem of computing reach-avoid probabilities for iterative predictions made with Bayesian neural network (BNN) models. Specifically, we leverage bound propagation techniques and backward recursion to compute lower bounds for the probability that trajectories of the BNN model reach a given set of states while avoiding a set of unsafe states. We use the lower bounds in the context of control and reinforcement learning to provide safety certification for given control policies, as well as to synthesize control policies that improve the certification bounds. On a set of benchmarks, we demonstrate that our framework can be employed to certify policies over BNNs predictions for problems of more than $10$ dimensions, and to effectively synthesize policies that significantly increase the lower bound on the satisfaction probability. △ Less

Submitted 19 June, 2021; v1 submitted 21 May, 2021; originally announced May 2021.

Comments: Accepted, UAI 2021. 17 pages

Adversarial Robustness Guarantees for Gaussian Processes

Authors: Andrea Patane, Arno Blaas, Luca Laurenti, Luca Cardelli, Stephen Roberts, Marta Kwiatkowska

Abstract: Gaussian processes (GPs) enable principled computation of model uncertainty, making them attractive for safety-critical applications. Such scenarios demand that GP decisions are not only accurate, but also robust to perturbations. In this paper we present a framework to analyse adversarial robustness of GPs, defined as invariance of the model's decision to bounded perturbations. Given a compact su… ▽ More Gaussian processes (GPs) enable principled computation of model uncertainty, making them attractive for safety-critical applications. Such scenarios demand that GP decisions are not only accurate, but also robust to perturbations. In this paper we present a framework to analyse adversarial robustness of GPs, defined as invariance of the model's decision to bounded perturbations. Given a compact subset of the input space $T\subseteq \mathbb{R}^d$, a point $x^*$ and a GP, we provide provable guarantees of adversarial robustness of the GP by computing lower and upper bounds on its prediction range in $T$. We develop a branch-and-bound scheme to refine the bounds and show, for any $ε> 0$, that our algorithm is guaranteed to converge to values $ε$-close to the actual values in finitely many iterations. The algorithm is anytime and can handle both regression and classification tasks, with analytical formulation for most kernels used in practice. We evaluate our methods on a collection of synthetic and standard benchmark datasets, including SPAM, MNIST and FashionMNIST. We study the effect of approximate inference techniques on robustness and demonstrate how our method can be used for interpretability. Our empirical results suggest that the adversarial robustness of GPs increases with accurate posterior estimation. △ Less

Submitted 7 April, 2021; originally announced April 2021.

Comments: Submitted for publication

Bayesian Inference with Certifiable Adversarial Robustness

Authors: Matthew Wicker, Luca Laurenti, Andrea Patane, Zhoutong Chen, Zheng Zhang, Marta Kwiatkowska

Abstract: We consider adversarial training of deep neural networks through the lens of Bayesian learning, and present a principled framework for adversarial training of Bayesian Neural Networks (BNNs) with certifiable guarantees. We rely on techniques from constraint relaxation of non-convex optimisation problems and modify the standard cross-entropy error model to enforce posterior robustness to worst-case… ▽ More We consider adversarial training of deep neural networks through the lens of Bayesian learning, and present a principled framework for adversarial training of Bayesian Neural Networks (BNNs) with certifiable guarantees. We rely on techniques from constraint relaxation of non-convex optimisation problems and modify the standard cross-entropy error model to enforce posterior robustness to worst-case perturbations in $ε$-balls around input points. We illustrate how the resulting framework can be combined with methods commonly employed for approximate inference of BNNs. In an empirical investigation, we demonstrate that the presented approach enables training of certifiably robust models on MNIST, FashionMNIST and CIFAR-10 and can also be beneficial for uncertainty calibration. Our method is the first to directly train certifiable BNNs, thus facilitating their deployment in safety-critical applications. △ Less

Submitted 22 February, 2021; v1 submitted 10 February, 2021; originally announced February 2021.

Comments: Accepted AISTATS2021

Resonance and antiresonance in Raman scattering in GaSe and InSe crystals

Authors: M. Osiekowicz, D. Staszczuk, K. Olkowska-Pucko, Ł. Kipczak, M. Grzeszczyk, M. Zinkiewicz, K. Nogajewski, Z. R. Kudrynskyi, Z. D. Kovalyuk, A. Patané, A. Babiński, M. R. Molas

Abstract: The temperature effect on the Raman scattering efficiency is investigated in $\varepsilon$-GaSe and $γ$-InSe crystals. We found that varying the temperature over a broad range from 5 K to 350 K permits to achieve both the resonant conditions and the antiresonance behaviour in Raman scattering of the studied materials. The resonant conditions of Raman scattering are observed at about 270 K under th… ▽ More The temperature effect on the Raman scattering efficiency is investigated in $\varepsilon$-GaSe and $γ$-InSe crystals. We found that varying the temperature over a broad range from 5 K to 350 K permits to achieve both the resonant conditions and the antiresonance behaviour in Raman scattering of the studied materials. The resonant conditions of Raman scattering are observed at about 270 K under the 1.96 eV excitation for GaSe due to the energy proximity of the optical band gap. In the case of InSe, the resonant Raman spectra are apparent at about 50 K and 270 K under correspondingly the 2.41 eV and 2.54 eV excitations as a result of the energy proximity of the \mbox{so-called} B transition. Interestingly, the observed resonances for both materials are followed by an antiresonance behaviour noticeable at higher temperatures than the detected resonances. The significant variations of phonon-modes intensities can be explained in terms of electron-phonon coupling and quantum interference of contributions from different points of the Brillouin zone △ Less

Submitted 24 October, 2020; originally announced October 2020.

Comments: 7 pages, 6 figures

Journal ref: Scientific Reports 11, 924 (2021)

Universal mobility characteristics of graphene originating from electron/hole scattering by ionised impurities

Authors: Jonathan H. Gosling, Oleg Makarovsky, Feiran Wang, Nathan D Cottam, Mark T. Greenaway, Amalia Patanè, Ricky Wildman, Christopher J. Tuck, Lyudmila Turyanska, T. Mark Fromhold

Abstract: Pristine graphene and graphene-based heterostructures exhibit exceptionally high electron mobility and conductance if their surface contains few electron-scattering impurities. Here, we reveal a universal connection between graphene's carrier mobility and the variation of its electrical conductance with carrier density. Our model of graphene conductivity is based on a convolution of carrier densit… ▽ More Pristine graphene and graphene-based heterostructures exhibit exceptionally high electron mobility and conductance if their surface contains few electron-scattering impurities. Here, we reveal a universal connection between graphene's carrier mobility and the variation of its electrical conductance with carrier density. Our model of graphene conductivity is based on a convolution of carrier density and its uncertainty, which reproduces the observed universality. Taking a single conductance measurement as input, this model accurately predicts the full shape of the conductance versus carrier density curves for a wide range of reported graphene samples. We verify the convolution model by numerically solving the Boltzmann transport equation to analyse in detail the effects of charged impurity scattering on carrier mobility. In this model, we also include optical phonons, which relax high-energy charge carriers for small impurity densities. Our numerical and analytical results both capture the universality observed in experiment and provide a way to estimate all key transport parameters of graphene devices. Our results demonstrate how the carrier mobility can be predicted and controlled, thereby providing insights for engineering the properties of 2D materials and heterostructures. △ Less

Submitted 16 May, 2020; originally announced May 2020.

Comments: 20 pages, 4 figures

MSC Class: 82

New Polymorphs of Two-Dimensional Indium Selenide with Enhanced Electronic Properties

Authors: Yuanhui Sun, Yawen Li, Tianshu Li, Koushik Biswas, Amalia Patanè, Lijun Zhang

Abstract: The two-dimensional (2D) semiconductor indium selenide (InSe) has attracted significant interest due its unique electronic band structure, high electron mobility and wide tunability of its band gap energy achieved by varying the layer thickness. All these features make 2D InSe a potential candidate for advanced electronic and optoelectronic applications. Here, we report on the discovery of new pol… ▽ More The two-dimensional (2D) semiconductor indium selenide (InSe) has attracted significant interest due its unique electronic band structure, high electron mobility and wide tunability of its band gap energy achieved by varying the layer thickness. All these features make 2D InSe a potential candidate for advanced electronic and optoelectronic applications. Here, we report on the discovery of new polymorphs of InSe with enhanced electronic properties. Using a global structure search that combines artificial swarm intelligence with first-principles energetic calculations, we identify polymorphs that consist of a centrosymmetric monolayer belonging to the point group D$_{3d}$, distinct from the well-known polymorphs based on the D$_{3h}$ monolayers that lack inversion symmetry. The new polymorphs are thermodynamically and kinetically stable, and exhibit a wider optical spectral response and larger electron mobilities compared to the known polymorphs. We discuss opportunities to synthesize these newly discovered polymorphs and viable routes to identify them by X-ray diffraction, Raman spectroscopy and second harmonic generation experiments. △ Less

Submitted 27 April, 2020; originally announced April 2020.

Comments: 20 pages, 5 figures

Probabilistic Safety for Bayesian Neural Networks

Authors: Matthew Wicker, Luca Laurenti, Andrea Patane, Marta Kwiatkowska

Abstract: We study probabilistic safety for Bayesian Neural Networks (BNNs) under adversarial input perturbations. Given a compact set of input points, $T \subseteq \mathbb{R}^m$, we study the probability w.r.t. the BNN posterior that all the points in $T$ are mapped to the same region $S$ in the output space. In particular, this can be used to evaluate the probability that a network sampled from the BNN is… ▽ More We study probabilistic safety for Bayesian Neural Networks (BNNs) under adversarial input perturbations. Given a compact set of input points, $T \subseteq \mathbb{R}^m$, we study the probability w.r.t. the BNN posterior that all the points in $T$ are mapped to the same region $S$ in the output space. In particular, this can be used to evaluate the probability that a network sampled from the BNN is vulnerable to adversarial attacks. We rely on relaxation techniques from non-convex optimization to develop a method for computing a lower bound on probabilistic safety for BNNs, deriving explicit procedures for the case of interval and linear function propagation techniques. We apply our methods to BNNs trained on a regression task, airborne collision avoidance, and MNIST, empirically showing that our approach allows one to certify probabilistic safety of BNNs with millions of parameters. △ Less

Submitted 18 June, 2020; v1 submitted 21 April, 2020; originally announced April 2020.

Comments: UAI 2020; 13 pages, 5 figures, 1 table

Photoluminescence dynamics in few-layer InSe

Authors: Tommaso Venanzi, Himani Arora, Stephan Winnerl, Alexej Pashkin, Phanish Chava, Amalia Patanè, Zakhar D. Kovalyuk, Zalhar R. Kudrynskyi, Kenji Watanabe, Takashi Taniguchi, Artur Erbe, Manfred Helm, Harald Schneider

Abstract: We study the optical properties of thin flakes of InSe encapsulated in hBN. More specifically, we investigate the photoluminescence (PL) emission and its dependence on sample thickness and temperature. Through the analysis of the PL lineshape, we discuss the relative weights of the exciton and electron-hole contributions. Thereafter we investigate the PL dynamics. Two contributions are distinguish… ▽ More We study the optical properties of thin flakes of InSe encapsulated in hBN. More specifically, we investigate the photoluminescence (PL) emission and its dependence on sample thickness and temperature. Through the analysis of the PL lineshape, we discuss the relative weights of the exciton and electron-hole contributions. Thereafter we investigate the PL dynamics. Two contributions are distinguishable at low temperature: direct bandgap electron-hole and defect-assisted recombination. The two recombination processes have lifetime of $τ_1 \sim 8\;$ns and $τ_2 \sim 100\;$ns, respectively. The relative weights of the direct bandgap and defect-assisted contributions show a strong layer dependence due to the direct-to-indirect bandgap crossover. Electron-hole PL lifetime is limited by population transfer to lower-energy states and no dependence on the number of layers was observed. The lifetime of the defect-assisted recombination gets longer for thinner samples. Finally, we show that the PL lifetime decreases at high temperatures as a consequence of more efficient non-radiative recombinations. △ Less

Submitted 31 March, 2020; v1 submitted 27 March, 2020; originally announced March 2020.

Comments: The manuscript will be published in the journal Physical Review Materials. Copyright 2011 by American Physical Society (https://journals.aps.org/prmaterials/)

Robustness of Bayesian Neural Networks to Gradient-Based Attacks

Authors: Ginevra Carbone, Matthew Wicker, Luca Laurenti, Andrea Patane, Luca Bortolussi, Guido Sanguinetti

Abstract: Vulnerability to adversarial attacks is one of the principal hurdles to the adoption of deep learning in safety-critical applications. Despite significant efforts, both practical and theoretical, the problem remains open. In this paper, we analyse the geometry of adversarial attacks in the large-data, overparametrized limit for Bayesian Neural Networks (BNNs). We show that, in the limit, vulnerabi… ▽ More Vulnerability to adversarial attacks is one of the principal hurdles to the adoption of deep learning in safety-critical applications. Despite significant efforts, both practical and theoretical, the problem remains open. In this paper, we analyse the geometry of adversarial attacks in the large-data, overparametrized limit for Bayesian Neural Networks (BNNs). We show that, in the limit, vulnerability to gradient-based attacks arises as a result of degeneracy in the data distribution, i.e., when the data lies on a lower-dimensional submanifold of the ambient space. As a direct consequence, we demonstrate that in the limit BNN posteriors are robust to gradient-based adversarial attacks. Experimental results on the MNIST and Fashion MNIST datasets with BNNs trained with Hamiltonian Monte Carlo and Variational Inference support this line of argument, showing that BNNs can display both high accuracy and robustness to gradient based adversarial attacks. △ Less

Submitted 24 June, 2020; v1 submitted 11 February, 2020; originally announced February 2020.

Interlayer band-to-band tunneling and negative differential resistance in van der Waals BP/InSe field effect transistors

Authors: Quanshan Lv, Faguang Yan, Nobuya Mori, Wenkai Zhu, Ce Hu, Zakhar R. Kudrynskyi, Zakhar D. Kovalyuk, Amalia Patanè, Kaiyou Wang

Abstract: Atomically thin layers of van der Waals (vdW) crystals offer an ideal material platform to realize tunnel field effect transistors (TFETs) that exploit the tunneling of charge carriers across the forbidden gap of a vdW heterojunction. This type of device requires a precise energy band alignment of the different layers of the junction to optimize the tunnel current. Amongst two-dimensional (2D) vdW… ▽ More Atomically thin layers of van der Waals (vdW) crystals offer an ideal material platform to realize tunnel field effect transistors (TFETs) that exploit the tunneling of charge carriers across the forbidden gap of a vdW heterojunction. This type of device requires a precise energy band alignment of the different layers of the junction to optimize the tunnel current. Amongst two-dimensional (2D) vdW materials, black phosphorus (BP) and indium selenide (InSe) have a Brillouin zone-centered conduction and valence bands, and a type II band offset, both ideally suited for band-to-band tunneling. Here, we demonstrate TFETs based on BP/InSe heterojunctions with diverse electrical transport characteristics: forward rectifying, Zener-tunneling and backward rectifying characteristics are realized in BP/InSe junctions with different thickness of the BP layer or by electrostatic gating of the junction. Electrostatic gating yields a large on/off current ratio of up to 108 and negative differential resistance at low applied voltages (V ~ 0.2V). These findings illustrate versatile functionalities of TFETs based on BP and InSe, offering opportunities for applications of these 2D materials beyond the device architectures reported in the current literature. △ Less

Submitted 28 January, 2020; originally announced January 2020.

Design of van der Waals Interfaces for Broad-Spectrum Optoelectronics

Authors: Nicolas Ubrig, Evgeniy Ponomarev, Johanna Zultak, Daniil Domaretskiy, Viktor Zólyomi, Daniel Terry, James Howarth, Ignacio Gutiérrez-Lezama, Alexander Zhukov, Zakhar R. Kudrynskyi, Zakhar D. Kovalyuk, Amalia Patanè, Takashi Taniguchi, Kenji Watanabe, Roman V. Gorbachev, Vladimir I. Fal'ko, Alberto F. Morpurgo

Abstract: Van der Waals (vdW) materials offer new ways to assemble artificial electronic media with properties controlled at the design stage, by combining atomically defined layers into interfaces and heterostructures. Their potential for optoelectronics stems from the possibility to tailor the spectral response over a broad range by exploiting interlayer transitions between different compounds with an app… ▽ More Van der Waals (vdW) materials offer new ways to assemble artificial electronic media with properties controlled at the design stage, by combining atomically defined layers into interfaces and heterostructures. Their potential for optoelectronics stems from the possibility to tailor the spectral response over a broad range by exploiting interlayer transitions between different compounds with an appropriate band-edge alignment. For the interlayer transitions to be radiative, however, a serious challenge comes from details of the materials --such as lattice mismatch or even a small misalignment of the constituent layers-- that can drastically suppress the electron-photon coupling. The problem was evidenced in recent studies of heterostructures of monolayer transition metal dichalcogenides, whose band edges are located at the K-point of reciprocal space. Here we demonstrate experimentally that the solution to the interlayer coupling problem is to engineer type-II interfaces by assembling atomically thin crystals that have the bottom of the conduction band and the top of the valence band at the $Γ$-point, thus avoiding any momentum mismatch. We find that this type of vdW interfaces exhibits radiative optical transition irrespective of lattice constant, rotational/translational alignment of the two layers, or whether the constituent materials are direct or indirect gap semiconductors. The result, which is robust and of general validity, drastically broadens the scope of future optoelectronics device applications based on 2D materials. △ Less

Submitted 4 February, 2020; v1 submitted 21 December, 2019; originally announced December 2019.

Journal ref: Nature Materials (2020)

Safety Guarantees for Planning Based on Iterative Gaussian Processes

Authors: Kyriakos Polymenakos, Luca Laurenti, Andrea Patane, Jan-Peter Calliess, Luca Cardelli, Marta Kwiatkowska, Alessandro Abate, Stephen Roberts

Abstract: Gaussian Processes (GPs) are widely employed in control and learning because of their principled treatment of uncertainty. However, tracking uncertainty for iterative, multi-step predictions in general leads to an analytically intractable problem. While approximation methods exist, they do not come with guarantees, making it difficult to estimate their reliability and to trust their predictions. I… ▽ More Gaussian Processes (GPs) are widely employed in control and learning because of their principled treatment of uncertainty. However, tracking uncertainty for iterative, multi-step predictions in general leads to an analytically intractable problem. While approximation methods exist, they do not come with guarantees, making it difficult to estimate their reliability and to trust their predictions. In this work, we derive formal probability error bounds for iterative prediction and planning with GPs. Building on GP properties, we bound the probability that random trajectories lie in specific regions around the predicted values. Namely, given a tolerance $ε> 0 $, we compute regions around the predicted trajectory values, such that GP trajectories are guaranteed to lie inside them with probability at least $1-ε$. We verify experimentally that our method tracks the predictive uncertainty correctly, even when current approximation techniques fail. Furthermore, we show how the proposed bounds can be employed within a safe reinforcement learning framework to verify the safety of candidate control policies, guiding the synthesis of provably safe controllers. △ Less

Submitted 7 September, 2020; v1 submitted 29 November, 2019; originally announced December 2019.

Comments: An earlier version of this work presented in NeurIPS-2019 Workshop on Safety and Robustness in Decision Making. A shorter (but otherwise equivalent) paper was accepted to the 59th Conference on Decision and Control (CDC2020)

Realization of Universal Quantum Gates with Spin-Qudits in Colloidal Quantum Dots

Authors: Fabrizio Moro, Alistair J. Fielding, Lyudmila Turyanska, Amalia Patanè

Abstract: We exploit hyperfine interactions in a single Mn-ion confined in a quantum dot (QD) to create a qudit, i.e. a multi-level quantum-bit system, with well defined, addressable and robust set of spin states for the realization of universal quantum gates. We generate and probe an arbitrary superposition of states between selected hyperfine energy level pairs by using electron double resonance detected… ▽ More We exploit hyperfine interactions in a single Mn-ion confined in a quantum dot (QD) to create a qudit, i.e. a multi-level quantum-bit system, with well defined, addressable and robust set of spin states for the realization of universal quantum gates. We generate and probe an arbitrary superposition of states between selected hyperfine energy level pairs by using electron double resonance detected nuclear magnetic resonance (EDNMR). This enables the observation of Rabi oscillations and the experimental realization of NOT and SWAP universal quantum gates that are robust against decoherence. Our protocol for cyclical preparation, manipulation and read-out of logic gates offers opportunities for integration of qudits in scalable quantum circuit architectures beyond solid state electron spin qubits. △ Less

Submitted 17 September, 2019; originally announced September 2019.

Two-dimensional covalent crystals by chemical conversion of thin van der Waals materials

Authors: Vishnu Sreepal, Mehmet Yagmurcukardes, Kalangi S. Vasu, Daniel J. Kelly, Sarah F. R. Taylor, Vasyl G. Kravets, Zakhar Kudrynskyi, Zakhar D. Kovalyuk, Amalia Patanè, Alexander N. Grigorenko, Sarah J. Haigh, Christopher Hardacre, Laurence Eaves, Hasan Sahin, Andre K. Geim, Francois M. Peeters, Rahul R. Nair

Abstract: Most of the studied two-dimensional (2D) materials have been obtained by exfoliation of van der Waals crystals. Recently, there has been growing interest in fabricating synthetic 2D crystals which have no layered bulk analogues. These efforts have been focused mainly on the surface growth of molecules in high vacuum. Here, we report an approach to making 2D crystals of covalent solids by chemical… ▽ More Most of the studied two-dimensional (2D) materials have been obtained by exfoliation of van der Waals crystals. Recently, there has been growing interest in fabricating synthetic 2D crystals which have no layered bulk analogues. These efforts have been focused mainly on the surface growth of molecules in high vacuum. Here, we report an approach to making 2D crystals of covalent solids by chemical conversion of van der Waals layers. As an example, we use 2D indium selenide (InSe) obtained by exfoliation and converted it by direct fluorination into indium fluoride (InF3), which has a non-layered, rhombohedral structure and therefore cannot be possibly obtained by exfoliation. The conversion of InSe into InF3 is found to be feasible for thicknesses down to three layers of InSe, and the obtained stable InF3 layers are doped with selenium. We study this new 2D material by optical, electron transport and Raman measurements and show that it is a semiconductor with a direct bandgap of 2.2 eV, exhibiting high optical transparency across the visible and infrared spectral ranges. We also demonstrate the scalability of our approach by chemical conversion of large-area, thin InSe laminates obtained by liquid exfoliation into InF3 films. The concept of chemical conversion of cleavable thin van der Waals crystals into covalently-bonded non-cleavable ones opens exciting prospects for synthesizing a wide variety of novel atomically thin covalent crystals. △ Less

Submitted 20 August, 2019; originally announced August 2019.

Journal ref: Nano Lett.2019, 19, 6475, 2019

Adversarial Robustness Guarantees for Classification with Gaussian Processes

Authors: Arno Blaas, Andrea Patane, Luca Laurenti, Luca Cardelli, Marta Kwiatkowska, Stephen Roberts

Abstract: We investigate adversarial robustness of Gaussian Process Classification (GPC) models. Given a compact subset of the input space $T\subseteq \mathbb{R}^d$ enclosing a test point $x^*$ and a GPC trained on a dataset $\mathcal{D}$, we aim to compute the minimum and the maximum classification probability for the GPC over all the points in $T$. In order to do so, we show how functions lower- and upper… ▽ More We investigate adversarial robustness of Gaussian Process Classification (GPC) models. Given a compact subset of the input space $T\subseteq \mathbb{R}^d$ enclosing a test point $x^*$ and a GPC trained on a dataset $\mathcal{D}$, we aim to compute the minimum and the maximum classification probability for the GPC over all the points in $T$. In order to do so, we show how functions lower- and upper-bounding the GPC output in $T$ can be derived, and implement those in a branch and bound optimisation algorithm. For any error threshold $ε> 0$ selected a priori, we show that our algorithm is guaranteed to reach values $ε$-close to the actual values in finitely many iterations. We apply our method to investigate the robustness of GPC models on a 2D synthetic dataset, the SPAM dataset and a subset of the MNIST dataset, providing comparisons of different GPC training techniques, and show how our method can be used for interpretability analysis. Our empirical analysis suggests that GPC robustness increases with more accurate posterior estimation. △ Less

Submitted 11 March, 2020; v1 submitted 28 May, 2019; originally announced May 2019.

Comments: 10 pages, 6 figures + Supplementary Material

Enhanced photoresponse in MoTe2 photodetectors with asymmetric graphene contacts

Authors: Xia Wei, Faguang Yan, Quanshan Lv, Wenkai Zhu, Ce Hu, Amalia Patan`e, Kaiyou Wang

Abstract: Atomically thin two dimensional (2D) materials are promising candidates for miniaturized high-performance optoelectronic devices. Here, we report on multilayer MoTe2 photodetectors contacted with asymmetric electrodes based on n- and p-type graphene layers. The asymmetry in the graphene contacts creates a large (Ebi ~100 kV cm-1) built-in electric field across the short (l = 15 nm) MoTe2 channel,… ▽ More Atomically thin two dimensional (2D) materials are promising candidates for miniaturized high-performance optoelectronic devices. Here, we report on multilayer MoTe2 photodetectors contacted with asymmetric electrodes based on n- and p-type graphene layers. The asymmetry in the graphene contacts creates a large (Ebi ~100 kV cm-1) built-in electric field across the short (l = 15 nm) MoTe2 channel, causing a high and broad (400 to 1400 nm) photoresponse even without any externally applied voltage. Spatially resolved photovoltage maps reveal an enhanced photoresponse and larger built-in electric field in regions of the MoTe2 layer between the two graphene contacts. Furthermore, a fast (~0.01 ms) photoresponse is achieved in both the photovoltaic and photoconductive operation modes of the junction. Our findings could be extended to other 2D materials and offer prospects for the implementation of asymmetric graphene contacts in future low-power optoelectronic applications. △ Less

Submitted 7 April, 2019; v1 submitted 21 March, 2019; originally announced March 2019.

Comments: 19 pages, 5 figures

Statistical Guarantees for the Robustness of Bayesian Neural Networks

Authors: Luca Cardelli, Marta Kwiatkowska, Luca Laurenti, Nicola Paoletti, Andrea Patane, Matthew Wicker

Abstract: We introduce a probabilistic robustness measure for Bayesian Neural Networks (BNNs), defined as the probability that, given a test point, there exists a point within a bounded set such that the BNN prediction differs between the two. Such a measure can be used, for instance, to quantify the probability of the existence of adversarial examples. Building on statistical verification techniques for pr… ▽ More We introduce a probabilistic robustness measure for Bayesian Neural Networks (BNNs), defined as the probability that, given a test point, there exists a point within a bounded set such that the BNN prediction differs between the two. Such a measure can be used, for instance, to quantify the probability of the existence of adversarial examples. Building on statistical verification techniques for probabilistic models, we develop a framework that allows us to estimate probabilistic robustness for a BNN with statistical guarantees, i.e., with a priori error and confidence bounds. We provide experimental comparison for several approximate BNN inference techniques on image classification tasks associated to MNIST and a two-class subset of the GTSRB dataset. Our results enable quantification of uncertainty of BNN predictions in adversarial settings. △ Less

Submitted 5 March, 2019; originally announced March 2019.

Comments: 9 pages, 6 figures

Tunnel spectroscopy of localised electronic states in hexagonal boron nitride

Authors: M. T. Greenaway, E. E. Vdovin, D. Ghazaryan, A. Misra, A. Mishchenko, Y. Cao, Z. Wang, J. R. Wallbank, M. Holwill, Yu. N. Khanin, S. V. Morozov, K. Watanabe, T. Taniguchi, O. Makarovsky, T. M. Fromhold, A. Patanè, A. K. Geim, V. I. Fal'ko, K. S. Novoselov, L. Eaves

Abstract: Hexagonal boron nitride (hBN) is a large band gap layered crystal, frequently incorporated in van der Waals (vdW) heterostructures as an insulating or tunnel barrier. Localised states with energies within its band gap can emit visible light, relevant to applications in nanophotonics and quantum information processing. However, they also give rise to conducting channels, which can induce electrical… ▽ More Hexagonal boron nitride (hBN) is a large band gap layered crystal, frequently incorporated in van der Waals (vdW) heterostructures as an insulating or tunnel barrier. Localised states with energies within its band gap can emit visible light, relevant to applications in nanophotonics and quantum information processing. However, they also give rise to conducting channels, which can induce electrical breakdown when a large voltage is applied. Here we use gated tunnel transistors to study resonant electron tunnelling through the localised states in few atomic-layer hBN barriers sandwiched between two monolayer graphene electrodes. The measurements are used to determine the energy, linewidth, tunnelling transmission probability, and depth within the barrier of more than 50 distinct localised states. A three-step process of electron percolation through two spatially separated localised states is also investigated. △ Less

Submitted 14 December, 2018; v1 submitted 2 October, 2018; originally announced October 2018.

Comments: Supplementary information can be found at https://doi.org/10.1038/s42005-018-0097-1

Journal ref: Communications Physics 1, 94 (2018)

Robustness Guarantees for Bayesian Inference with Gaussian Processes

Authors: Luca Cardelli, Marta Kwiatkowska, Luca Laurenti, Andrea Patane

Abstract: Bayesian inference and Gaussian processes are widely used in applications ranging from robotics and control to biological systems. Many of these applications are safety-critical and require a characterization of the uncertainty associated with the learning model and formal guarantees on its predictions. In this paper we define a robustness measure for Bayesian inference against input perturbations… ▽ More Bayesian inference and Gaussian processes are widely used in applications ranging from robotics and control to biological systems. Many of these applications are safety-critical and require a characterization of the uncertainty associated with the learning model and formal guarantees on its predictions. In this paper we define a robustness measure for Bayesian inference against input perturbations, given by the probability that, for a test point and a compact set in the input space containing the test point, the prediction of the learning model will remain $δ-$close for all the points in the set, for $δ>0.$ Such measures can be used to provide formal guarantees for the absence of adversarial examples. By employing the theory of Gaussian processes, we derive tight upper bounds on the resulting robustness by utilising the Borell-TIS inequality, and propose algorithms for their computation. We evaluate our techniques on two examples, a GP regression problem and a fully-connected deep neural network, where we rely on weak convergence to GPs to study adversarial examples on the MNIST dataset. △ Less

Submitted 24 October, 2018; v1 submitted 17 September, 2018; originally announced September 2018.

Magnetotransport and lateral confinement in an InSe van der Waals Heterostructure

Authors: Yong** Lee, Riccardo Pisoni, Hiske Overweg, Marius Eich, Peter Rickhaus, Amalia Patanè, Zakhar R. Kudrynskyi, Zakhar. D. Kovalyuk, Roman Gorbachev, Kenji Watanabe, Takashi Taniguchi, Thomas Ihn, Klaus Ensslin

Abstract: In the last six years, Indium selenide (InSe) has appeared as a new van der Waals heterostructure platform which has been extensively studied due to its unique electronic and optical properties. Such as transition metal dichalcogenides (TMDCs), the considerable bandgap and high electron mobility can provide a potential optoelectronic application. Here we present low-temperature transport measureme… ▽ More In the last six years, Indium selenide (InSe) has appeared as a new van der Waals heterostructure platform which has been extensively studied due to its unique electronic and optical properties. Such as transition metal dichalcogenides (TMDCs), the considerable bandgap and high electron mobility can provide a potential optoelectronic application. Here we present low-temperature transport measurements on a few-layer InSe van der Waals heterostructure with graphene-gated contacts. For high magnetic fields, we observe magnetoresistance minima at even filling factors related to two-fold spin degeneracy. By electrostatic gating with negatively biased split gates, a one-dimensional channel is realized. Close to pinch-off, transport through the constriction is dominated by localized states with charging energies ranging from 2 to 5 meV. This work opens new possibility to explore the low-dimensional physics including quantum point contact and quantum dot. △ Less

Submitted 28 May, 2018; originally announced May 2018.

Comments: 11 pages, 4 figures

Journal ref: 2D Mater. 5 035040,2018

Gate-Defined Quantum Confinement in InSe-based van der Waals Heterostructures

Authors: Matthew Hamer, Endre Tóvári, Mengjian Zhu, Michael D. Thompson, Alexander Mayorov, Jonathon Prance, Yong** Lee, Richard P. Haley, Zakhar R. Kudrynskyi, Amalia Patanè, Daniel Terry, Zakhar D. Kovalyuk, Klaus Ensslin, Andrey V. Kretinin, Andre Geim, Roman Gorbachev

Abstract: Indium selenide, a post-transition metal chalcogenide, is a novel two-dimensional (2D) semiconductor with interesting electronic properties. Its tunable band gap and high electron mobility have already attracted considerable research interest. Here we demonstrate strong quantum confinement and manipulation of single electrons in devices made from few-layer crystals of InSe using electrostatic gati… ▽ More Indium selenide, a post-transition metal chalcogenide, is a novel two-dimensional (2D) semiconductor with interesting electronic properties. Its tunable band gap and high electron mobility have already attracted considerable research interest. Here we demonstrate strong quantum confinement and manipulation of single electrons in devices made from few-layer crystals of InSe using electrostatic gating. We report on gate-controlled quantum dots in the Coulomb blockade regime as well as one-dimensional quantization in point contacts, revealing multiple plateaus. The work represents an important milestone in the development of quality devices based on 2D materials and makes InSe a prime candidate for relevant electronic and optoelectronic applications. △ Less

Submitted 15 May, 2018; originally announced May 2018.

Comments: 13 pages, 3 figures

Journal ref: Nano Lett., 2018, 18 (6), p 3950

Fast multicolor photodetectors based on graphene-contacted p-GaSe/n-InSe van der Waals heterostructures

Authors: Faguang Yan, Lixia Zhao, Amalia Patanè, **An Hu, Xia Wei, Wengang Luo, Dong Zhang, Quanshan Lv, Qi Feng, Chao Shen, Kai Chang, Laurence Eaves, Kaiyou Wang

Abstract: The integration of different two-dimensional materials within a multilayer van der Waals (vdW) heterostructure offers a promising technology for realizing high performance opto-electronic devices such as photodetectors and light sources1-3. Transition metal dichalcogenides, e.g. MoS2 and WSe2, have been employed as the optically-active layer in recently developed heterojunctions. However, MoS2 and… ▽ More The integration of different two-dimensional materials within a multilayer van der Waals (vdW) heterostructure offers a promising technology for realizing high performance opto-electronic devices such as photodetectors and light sources1-3. Transition metal dichalcogenides, e.g. MoS2 and WSe2, have been employed as the optically-active layer in recently developed heterojunctions. However, MoS2 and WSe2 become direct band gap semiconductors only in mono- or bilayer form4,5. In contrast, the metal monochalcogenides InSe and GaSe retain a direct bandgap over a wide range of layer thicknesses from bulk crystals down to exfoliated flakes only a few atomic monolayers thick6,7. Here we report on vdW heterojunction diodes based on InSe and GaSe: the type II band alignment between the two materials and their distinctive spectral response, combined with the low electrical resistance of transparent graphene electrodes, enable effective separation and extraction of photoexcited carriers from the heterostructure even when no external voltage is applied. Our devices are fast (< 10 μs), self-driven photodetectors with multicolor photoresponse ranging from the ultraviolet to the near-infrared and have the potential to accelerate the exploitation of two-dimensional vdW crystals by creating new routes to miniaturized optoelectronics beyond present technologies. △ Less

Submitted 6 March, 2017; originally announced March 2017.

High Electron Mobility, Quantum Hall Effect and Anomalous Optical Response in Atomically Thin InSe

Authors: D. A. Bandurin, A. V. Tyurnina, G. L. Yu, A. Mishchenko, V. Zolyomi, S. V. Morozov, R. Krishna Kumar, R. V. Gorbachev, Z. R. Kudrynskyi, S. Pezzini, Z. D. Kovalyuk, U. Zeitler, K. S. Novoselov, A. Patane, L. Eaves, I. V. Grigorieva, V. I. Fal'ko, A. K. Geim, Y. Cao

Abstract: A decade of intense research on two-dimensional (2D) atomic crystals has revealed that their properties can differ greatly from those of the parent compound. These differences are governed by changes in the band structure due to quantum confinement and are most profound if the underlying lattice symmetry changes. Here we report a high-quality 2D electron gas in few-layer InSe encapsulated in hexag… ▽ More A decade of intense research on two-dimensional (2D) atomic crystals has revealed that their properties can differ greatly from those of the parent compound. These differences are governed by changes in the band structure due to quantum confinement and are most profound if the underlying lattice symmetry changes. Here we report a high-quality 2D electron gas in few-layer InSe encapsulated in hexagonal boron nitride under an inert atmosphere. Carrier mobilities are found to exceed 1,000 and 10,000 cm2/Vs at room and liquid-helium temperatures, respectively, allowing the observation of the fully-developed quantum Hall effect. The conduction electrons occupy a single 2D subband and have a small effective mass. Photoluminescence spectroscopy reveals that the bandgap increases by more than 0.5 eV with decreasing the thickness from bulk to bilayer InSe. The band-edge optical response vanishes in monolayer InSe, which is attributed to monolayer's mirror-plane symmetry. Encapsulated 2D InSe expands the family of graphene-like semiconductors and, in terms of quality, is competitive with atomically-thin dichalcogenides and black phosphorus. △ Less

Submitted 31 August, 2016; originally announced August 2016.

Journal ref: Nature Nanotechnology 12, 223-227 (2016)

Phonon-assisted resonant tunneling of electrons in graphene-boron nitride transistors

Authors: E. E. Vdovin, A. Mishchenko, M. T. Greenaway, M. J. Zhu, D. Ghazaryan, A. Misra, Y. Cao, S. V. Morozov, O. Makarovsky, A. Patanè, G. J. Slotman, M. I. Katsnelson, A. K. Geim, K. S. Novoselov, L. Eaves

Abstract: We observe a series of sharp resonant features in the differential conductance of graphene-hexagonal boron nitride-graphene tunnel transistors over a wide range of bias voltages between $\sim$10 and 200 mV. We attribute them to electron tunneling assisted by the emission of phonons of well-defined energy. The bias voltages at which they occur are insensitive to the applied gate voltage and hence i… ▽ More We observe a series of sharp resonant features in the differential conductance of graphene-hexagonal boron nitride-graphene tunnel transistors over a wide range of bias voltages between $\sim$10 and 200 mV. We attribute them to electron tunneling assisted by the emission of phonons of well-defined energy. The bias voltages at which they occur are insensitive to the applied gate voltage and hence independent of the carrier densities in the graphene electrodes, so plasmonic effects can be ruled out. The phonon energies corresponding to the resonances are compared with the lattice dispersion curves of graphene-boron nitride heterostructures and are close to peaks in the single phonon density of states. △ Less

Submitted 8 December, 2015; v1 submitted 7 December, 2015; originally announced December 2015.

Journal ref: Phys. Rev. Lett. 116, 186603 (2016)

Resonant tunnelling between the chiral Landau states of twisted graphene lattices

Authors: M. T. Greenaway, E. E. Vdovin, A. Mishchenko, O. Makarovsky, A. Patanè, J. R. Wallbank, Y. Cao, A. V. Kretinin, M. J. Zhu, S. V. Morozov, V. I. Fal'ko, K. S. Novoselov, A. K. Geim, T. M. Fromhold, L. Eaves

Abstract: A new class of multilayered functional materials has recently emerged in which the component atomic layers are held together by weak van der Waals forces that preserve the structural integrity and physical properties of each layer. An exemplar of such a structure is a transistor device in which relativistic Dirac Fermions can resonantly tunnel through a boron nitride barrier, a few atomic layers t… ▽ More A new class of multilayered functional materials has recently emerged in which the component atomic layers are held together by weak van der Waals forces that preserve the structural integrity and physical properties of each layer. An exemplar of such a structure is a transistor device in which relativistic Dirac Fermions can resonantly tunnel through a boron nitride barrier, a few atomic layers thick, sandwiched between two graphene electrodes. An applied magnetic field quantises graphene's gapless conduction and valence band states into discrete Landau levels, allowing us to resolve individual inter-Landau level transitions and thereby demonstrate that the energy, momentum and chiral properties of the electrons are conserved in the tunnelling process. We also demonstrate that the change in the semiclassical cyclotron trajectories, following a tunnelling event, is a form of Klein tunnelling for inter-layer transitions. △ Less

Submitted 21 October, 2015; v1 submitted 21 September, 2015; originally announced September 2015.

Comments: in Nature Physics (2015)

Journal ref: Nature Physics 11, 1057-1062 (2015)

High Curie temperatures at low compensation in the ferromagnetic semiconductor (Ga,Mn)As

Authors: M. Wang, K. W. Edmonds, B. L. Gallagher, A. W. Rushforth, O. Makarovsky, A. Patanè, R. P. Campion, C. T. Foxon, V. Novak, T. Jungwirth

Abstract: We investigate the relationship between the Curie temperature TC and the carrier density p in the ferromagnetic semiconductor (Ga,Mn)As. Carrier densities are extracted from analysis of the Hall resistance at low temperatures and high magnetic fields. Results are found to be consistent with ion channeling measurements when performed on the same samples. We find that both TC and the electrical cond… ▽ More We investigate the relationship between the Curie temperature TC and the carrier density p in the ferromagnetic semiconductor (Ga,Mn)As. Carrier densities are extracted from analysis of the Hall resistance at low temperatures and high magnetic fields. Results are found to be consistent with ion channeling measurements when performed on the same samples. We find that both TC and the electrical conductivity increase monotonically with increasing p, and take their largest values when p is comparable to the concentration of substitutional Mn acceptors. This is inconsistent with models in which the Fermi level is located within a narrow isolated impurity band. △ Less

Submitted 11 March, 2013; originally announced March 2013.

Comments: 10 pages, 4 figures

Journal ref: Physical Review B 87, 121301 (2013)

Correspondence on "Controlling the Curie temperature in (Ga,Mn)As through location of the Fermi level within the impurity band" by M. Dobrowolska et al

Authors: K. W. Edmonds, B. L. Gallagher, M. Wang, A. W. Rushforth, O. Makarovsky, A. Patane, R. P. Campion, C. T. Foxon, V. Novak, T. Jungwirth

Abstract: Comment on the recent Nature Materials article by M. Dobrowolska et al., arXiv:1203.1852. We present experimental data showing that the Curie temperature and conductivity of high quality (Ga,Mn)As samples are maximized at low compensation, and thus the magnetic order in (Ga,Mn)As is not consistent with the isolated impurity band scenario. Comment on the recent Nature Materials article by M. Dobrowolska et al., arXiv:1203.1852. We present experimental data showing that the Curie temperature and conductivity of high quality (Ga,Mn)As samples are maximized at low compensation, and thus the magnetic order in (Ga,Mn)As is not consistent with the isolated impurity band scenario. △ Less

Submitted 16 November, 2012; originally announced November 2012.

Comments: 6 pages, 2 figures

Controlling high-frequency collective electron dynamics via single-particle complexity

Authors: N. Alexeeva, M. T. Greenaway, A. G. Balanov, O. Makarovsky, A. Patanè, M. B. Gaifullin, F. Kusmartsev, T. M. Fromhold

Abstract: We demonstrate, through experiment and theory, enhanced high-frequency current oscillations due to magnetically-induced conduction resonances in superlattices. Strong increase in the ac power originates from complex single-electron dynamics, characterized by abrupt resonant transitions between unbound and localized trajectories, which trigger and shape propagating charge domains. Our data demonstr… ▽ More We demonstrate, through experiment and theory, enhanced high-frequency current oscillations due to magnetically-induced conduction resonances in superlattices. Strong increase in the ac power originates from complex single-electron dynamics, characterized by abrupt resonant transitions between unbound and localized trajectories, which trigger and shape propagating charge domains. Our data demonstrate that external fields can tune the collective behavior of quantum particles by imprinting configurable patterns in the single-particle classical phase space. △ Less

Submitted 21 July, 2012; originally announced July 2012.

Comments: 5 pages, 4 figures

Journal ref: Phys. Rev. Lett. 109, 024102 (2012)

Optical imaging of resonant electrical carrier injection into individual quantum dots

Authors: A. Baumgartner, E. Stock, A. Patanè, L. Eaves, M. Henini, D. Bimberg

Abstract: We image the micro-electroluminescence (EL) spectra of self-assembled InAs quantum dots (QDs) embedded in the intrinsic region of a GaAs p-i-n diode and demonstrate optical detection of resonant carrier injection into a single QD. Resonant tunneling of electrons and holes into the QDs at bias voltages below the flat-band condition leads to sharp EL lines characteristic of individual QDs, accompani… ▽ More We image the micro-electroluminescence (EL) spectra of self-assembled InAs quantum dots (QDs) embedded in the intrinsic region of a GaAs p-i-n diode and demonstrate optical detection of resonant carrier injection into a single QD. Resonant tunneling of electrons and holes into the QDs at bias voltages below the flat-band condition leads to sharp EL lines characteristic of individual QDs, accompanied by a spatial fragmentation of the surface EL emission into small and discrete light- emitting areas, each with its own spectral fingerprint and Stark shift. We explain this behavior in terms of Coulomb interaction effects and the selective excitation of a small number of QDs within the ensemble due to preferential resonant tunneling paths for carriers. △ Less

Submitted 16 July, 2010; originally announced July 2010.

Comments: 4 pages

Journal ref: Phys. Rev. Lett. 105, 257401 (2010)

Sensitive detection of photoexcited carriers by resonant tunneling through a single quantum dot

Authors: E. E. Vdovin, O. Makarovsky, A. Patane, L. Eaves, Yu. N. Khanin

Abstract: We show that the resonant tunnel current through a single energy level of an individual quantum dot within an ensemble of dots is strongly sensitive to photoexcited holes that become bound in the close vicinity of the dot. The presence of these holes lowers the electrostatic energy of the quantum dot state and switches the current carrying channel from fully open to fully closed with a high on/o… ▽ More We show that the resonant tunnel current through a single energy level of an individual quantum dot within an ensemble of dots is strongly sensitive to photoexcited holes that become bound in the close vicinity of the dot. The presence of these holes lowers the electrostatic energy of the quantum dot state and switches the current carrying channel from fully open to fully closed with a high on/off ratio (> 50). The device can be reset by means of a bias voltage pulse. These properties are of interest for charge sensitive photon counting devices. △ Less

Submitted 9 July, 2009; originally announced July 2009.

Comments: 5 pages, 4 figures

Journal ref: Physical Review B 79 (2009) 193311

Bifurcations and chaos in semiconductor superlattices with a tilted magnetic field

Authors: A. G. Balanov, D. Fowler, A. Patanè, L. Eaves, T. M. Fromhold

Abstract: We study the effects of dissipation on electron transport in a semiconductor superlattice with an applied bias voltage and a magnetic field that is tilted relative to the superlattice axis.In previous work, we showed that although the applied fields are stationary,they act like a THz plane wave, which strongly couples the Bloch and cyclotron motion of electrons within the lowest miniband. As a c… ▽ More We study the effects of dissipation on electron transport in a semiconductor superlattice with an applied bias voltage and a magnetic field that is tilted relative to the superlattice axis.In previous work, we showed that although the applied fields are stationary,they act like a THz plane wave, which strongly couples the Bloch and cyclotron motion of electrons within the lowest miniband. As a consequence,the electrons exhibit a unique type of Hamiltonian chaos, which creates an intricate mesh of conduction channels (a stochastic web) in phase space, leading to a large resonant increase in the current flow at critical values of the applied voltage. This phase-space patterning provides a sensitive mechanism for controlling electrical resistance. In this paper, we investigate the effects of dissipation on the electron dynamics by modifying the semiclassical equations of motion to include a linear dam** term. We demonstrate that even in the presence of dissipation,deterministic chaos plays an important role in the electron transport process. We identify mechanisms for the onset of chaos and explore the associated sequence of bifurcations in the electron trajectories. When the Bloch and cyclotron frequencies are commensurate, complex multistability phenomena occur in the system. In particular, for fixed values of the control parameters several distinct stable regimes can coexist, each corresponding to different initial conditions. We show that this multistability has clear, experimentally-observable, signatures in the electron transport characteristics. △ Less

Submitted 30 October, 2007; v1 submitted 13 June, 2007; originally announced June 2007.

Comments: 14 pages 11 figures

Journal ref: Phys. Rev. E 77, 026209 (2008)

Anisotropic magnetic field dependence of many-body enhanced electron tunnelling through a quantum dot

Authors: E. E. Vdovin, Yu. N. Khanin, O. Makarovsky, A. Patane, L. Eaves, M. Henini, C. J. Mellor, K. A. Benedict, R. Airey

Abstract: We investigate the effect of an applied magnetic field on resonant tunneling of electrons through the bound states of self-assembled InAs quantum dots (QDs) embedded within an (AlGa)As tunnel barrier. At low temperatures (no more than 2 K), a magnetic field B applied either parallel or perpendicular to the direction of current flow causes a significant enhancement of the tunnel current. For the… ▽ More We investigate the effect of an applied magnetic field on resonant tunneling of electrons through the bound states of self-assembled InAs quantum dots (QDs) embedded within an (AlGa)As tunnel barrier. At low temperatures (no more than 2 K), a magnetic field B applied either parallel or perpendicular to the direction of current flow causes a significant enhancement of the tunnel current. For the latter field configuration, we observe a strong angular anisotropy of the enhanced current when B is rotated in the plane of the quantum dot layer. We attribute this behavior to the effect of the lowered symmetry of the QD eigenfunctions on the electron-electron interaction. △ Less

Submitted 23 March, 2007; originally announced March 2007.

Comments: Revtex4, 6 pages, 6 figures

Spatial map** of the electron eigenfunctions in InAs self-assembled quantum dots by magnetotunneling

Authors: E. E. Vdovin, Yu. N. Khanin, A. V. Veretennikov, A. Levin, A. Patane, Yu. V. Dubrovskii, L. Eaves, P. C. Main, M. Henini, G. Hill

Abstract: We use magnetotunnelling spectroscopy as a non-invasive probe to produce two-dimensional spatial images of the probability density of an electron confined in a self-assembled semiconductor quantum dot. The images reveal clearly the elliptical symmetry of the ground state and the characteristic lobes of the higher energy states. We use magnetotunnelling spectroscopy as a non-invasive probe to produce two-dimensional spatial images of the probability density of an electron confined in a self-assembled semiconductor quantum dot. The images reveal clearly the elliptical symmetry of the ground state and the characteristic lobes of the higher energy states. △ Less

Submitted 29 June, 2001; originally announced June 2001.

Comments: 5 pages, 5 figures, revtex

Stokes shift in quantum wells: trap** versus thermalization

Authors: A. Polimeni, A. Patanè, M. Grassi Alessi, M. Capizzi, F. Martelli, A. Bosacchi, S. Franchi

Abstract: Low temperature photoluminescence and photoluminescence excitation measurements have been performed in a set of InGaAs/GaAs samples with different indium molar fraction, well width, growth conditions and post-growth treatment. This has allowed to change in a controlled way the degree and source of disorder in the samples, thus resulting in an excitonic absorption linewidth varying between 1 and… ▽ More Low temperature photoluminescence and photoluminescence excitation measurements have been performed in a set of InGaAs/GaAs samples with different indium molar fraction, well width, growth conditions and post-growth treatment. This has allowed to change in a controlled way the degree and source of disorder in the samples, thus resulting in an excitonic absorption linewidth varying between 1 and 18 meV, and an ensueing Stokes shift changing between zero and 8 meV. The conditions of validity of two different models relating the Stokes shift to the linewidth broadening have been established in terms of different regimes of disorder and temperature. A continuous transition between those regimes has been demonstrated. △ Less

Submitted 17 September, 1996; originally announced September 1996.

Comments: File apstex, 12 p. + 4 Figures, to appear on Phys. Rev. B (Brief Reports.), 15 Dec. 1996. The figures will be faxed upon request. E-mail:[email protected] Fax: +39-6-4957697